Advanced search

On other websites (7429)

• Compiling FFMPEG on CentOS DigitalOcean

  29 July 2015, by coder_uk

  I set up a DigitalOcean instance running CentOS 6.5 and successfully followed the guide to compile FFMPEG (https://trac.ffmpeg.org/wiki/CompilationGuide/Centos). Hurrah!

  But of course I realised that by default, DigitalOcean creates a root user and so ffmpeg now lives in /root/bin/ffmpeg. Which isn’t ideal because when I want to exec the ffmpeg bin from nginx, I would have to run nginx as root for it to have permission.

  Questions ...

  1) Long-shot, but presumably if I change the owner of the ffmpeg binary to nginx, it still won’t work, because nginx won’t be able to access the /root folder it is in. Correct?

  2) I could run nginx as root (’user root’). But this seems like a very bad idea. Correct?

  3) Which leaves me with the option of creating a new user, and then compiling ffmpeg into its home folder. But: which user? EC2 creates ’ec2-user’, so should I make my own equivalent for DO? But then won’t I have to run nginx as that user, else I’ll run into the same problem?

  Or should I compile ffmpeg into the ’nginx’ home folder, if indeed it has one? Is that how it is supposed to be done?

  Since compiling ffmpeg takes ages, I don’t want to keep doing it, and the static files all seem very out of date. Thanks

• Adding a text overlay with constant size to a video with varying resolutions using ffmpeg

  22 December 2020, by BetaLixT

  I have a video with different resolutions at different points of the video and need to add a text overlay to it. I'm new with ffmpeg, the following is one of the commands I've tried

  


  ffmpeg -i input.mp4 -vf "drawtext=text='Hello World'" -c:a copy output.mp4


  


  But the output has blurry text with it resizing every few seconds (I'm guessing because of the various resolutions in the video) and some parts of the video have their aspect ratios altered.

  


  Information on input file

  


  The input.mp4 video is generated from some .h264 raw data converted to mp4 and some blank mp4 generated that are all concatenated together, following are the commands

  


  h264 to mp4:

  


  ffmpeg a.h264 -c copy a.mp4


  


  blank mp4

  


  ffmpeg -t 5 -f lavfi -i color=c=black:s=200x200 -c:v libx264 -tune stillimage -pix_fmt yuv420p -video_track_timescale 1200000 b.mp4


  


  Concat:

  


  ffmpeg -safe 0 -f concat -i list.txt -s 720x960 -c copy input.mp4


  


  I'd be thankful for any advice on how I could achieve this

  


• Ffmpeg - Generate VTT File From Sprite, Using Spatial Media Fragment

  20 October 2019, by David

  Hi I am looking to create a .VTT file from a sprite that i have generated using Ffmpeg.

  Ffmpeg command:

   $"-i {inputMediaFile} -vf \"select = not(mod(n\\, 30)),scale = 120:80,tile = 7x7\" -an -vsync 0 {outputMediaFile}"

  This selects every 30th frame, and then scales it to 120x80 pixels and creates 8x8 tiles in the output image.

  I would like to make a .VTT from the generated image in C#, so i know the height and width of my individual images in the sprite (120x80) and there is 64 images in total in the output image.

  From this i need to produce a VTT like this:

  WEBVTT
  
  
  
  1
  
  00:00:00.000 --> 00:00:01.000
  
  test-00001.jpg#xywh=0,0,120,80
  
  
  
  2
  
  00:00:01.000 --> 00:00:02.000
  
  test-00001.jpg#xywh=120,0,120,80
  
  
  
  3
  
  00:00:02.000 --> 00:00:03.000
  
  test-00001.jpg#xywh=240,0,120,80
  
  
  
  4
  
  00:00:03.000 --> 00:00:04.000
  
  test-00001.jpg#xywh=360,0,120,80
  
  
  
  5
  
  00:00:04.000 --> 00:00:05.000
  
  test-00001.jpg#xywh=480,0,120,80
  
  
  
  6
  
  00:00:05.000 --> 00:00:06.000
  
  test-00001.jpg#xywh=600,0,120,80
  
  
  
  7
  
  00:00:06.000 --> 00:00:07.000
  
  test-00001.jpg#xywh=720,0,120,80
  
  
  
  8
  
  00:00:07.000 --> 00:00:08.000
  
  test-00001.jpg#xywh=840,0,120,80
  
  
  
  9
  
  00:00:08.000 --> 00:00:09.000
  
  test-00001.jpg#xywh=0,80,120,80

  There is also situations when there is n amount of sprite files.

  Im hoping there may be a library out there that can handle this, or even better if i can keep it contained within Ffmpeg - based on Ffmpeg docs i dont think this is possible though.

  Thanks in advance if anyone as any ideas, its doable as ive seen Nodejs and Ruby examples.